RESUMEN
OBJECTIVE: To compare the sensitivity and specificity of a routine assessment (clinical impression) with a structured assessment which uses a validated assessment scale for the recognition of delirium on the intensive care unit (ICU). DESIGN: Observational study. METHOD: During their admission to the intensive care unit, 103 patients were assessed daily (with a maximum of 40 days) for the presence of delirium using the Confusion assessment method for the intensive care unit (CAM-ICU). Their physicians indicated whether or not they considered the patient delirious. These findings were compared. For all patients daily information was also collected about fixation and complications, such as self-extubation or self-removal of catheters. RESULTS: The patients were assessed for a period of 502 patient-days. CAM-ICU scores were positive (n = 108), negative (n = 235) or non-assessable because the patient was comatosed or deeply sedated (n = 159). The sensitivity of clinical detection by the physicians was 45% in comparison to the CAM-ICU. The specificity was high (97%). CONCLUSION: The diagnosis delirium is frequently missed on the ICU when only based on clinical impression. Routine assessment using a validated assessment scale such as the CAM-ICU might possibly improve this.